Preprint
Aug 2026
Self-Normalized Inference for Constant-Stepsize Temporal-Difference Learning under Markovian Sampling
A Brownian-bridge self-normalizer yields asymptotically pivotal confidence regions for prespecified state-value contrasts for prespecified state-value contrasts without estimating the long-run covariance or selecting a bandwidth or batch length.
Minmin Zeng, Yichen Zhang, Xiao-Feng Shao
· 0 citations